The present disclosure generally relates to semiconductor ion implantation, and more specifically, to methods for selectively implanting ionized carbon species into semiconductor work pieces.
Ion implanters can be used to treat silicon wafers by bombardment of the wafers with an ion beam. One use of such beam treatment is to selectively implant the wafers with impurities and/or dopants of a controlled concentration for fabrication of integrated circuits.
A typical ion implanter includes an ion source, an ion extraction device, a mass analysis device, a beam transport device and a wafer processing device. The ion source generates ions of desired atomic or molecular species. These ions are extracted from the source by an extraction system, typically a set of electrodes that energize and direct the flow of ions from the source. The desired ions are separated from byproducts of the ion source in a mass analysis device, typically a magnetic dipole performing mass dispersion of the extracted ion beam. The beam transport device, typically a vacuum system containing an optical train of focusing devices transports the ion beam to the wafer processing device while maintaining desired optical properties of the ion beam. Finally, the semiconductor wafers are implanted with the atomic or molecular species or ionic fragments thereof in the wafer processing device.
Ion energy is used to control junction depth in semiconductor devices. The energy of the ions that make up the ion beam determines the degree of depth of the implanted ions. High energy processes such as those used to form retrograde wells in semiconductor devices require implants of up to a few million electron-volts (MeV), while shallow junctions may only demand energies below 1 thousand electron-volts (keV), and ultra-shallow junctions may require energies as low as 250 electron-volts (eV). Typically, high current implanters (generally greater than 10 milliamps (mA) ion beam current) are used for high dose implants, while medium current implanters (generally capable up to about 1 mA beam current) are used for lower dose applications.
During semiconductor manufacturing it is sometimes beneficial to perform carbon implants. Specifically, it may be desirable to have carbon implants that yield a large amount of damage to the crystalline structure of the silicon substrate, for example, in which the integrated circuit is built. In some applications, it is beneficial to implant ionic carbon species to pre-amorphize selected portions of the silicon substrate. The ionized carbon implantation can be controlled so that only a certain depth of the silicon substrate is amorphized. The remaining depth remains crystalline. A dopant source such as phosphorus, arsenic, boron, or indium can then be implanted into a region of the amorphous portion and subsequently annealed to actively diffuse the dopant ions and recrystallize the amorphous portion. In this manner, the carbon species can provide pre-amorphization implantation and serves as a barrier to dopant diffusion for advanced source drain extension formation.
In other applications, it is sometimes beneficial to implant carbon to form p-type drain extensions so as to reduce transient enhanced diffusion of the p-dopant ion, e.g., boron. In addition, carbon ion implantation is sometimes beneficial during formation of shallow junctions so as to provide control over recrystallization of the amorphous regions during a subsequent annealing step. Carbon is also being investigated in conjunction with providing strain in NMOS implantation.
Presently, the most common precursor for carbon implants is carbon dioxide (CO2). One of the problems associated with the use of CO2 is the reduced operating lifetime of the ion source. During generation of the carbon ions from CO2 with the ion source, 2 oxygen atoms are also generated. These oxygen atoms are highly reactive (second in reactivity only to fluorine) and can deleteriously corrode, i.e., oxidize, the ion source, which is typically a metal electrode, thereby decreasing the operating lifetime.
Phosphine gases have been known to be added to carbon dioxide as a co-gas so as to react with the free oxygen generated by ionization of carbon dioxide rendering it less destructive to the ion source. However, as a co-gas, phosphine adds gas flow and pressure to the ion source without adding any usable precursor material. Moreover, phosphine is relative toxic and expensive
It would be desirable to minimize oxidation of the ion sources when utilized with carbon dioxide for generation of carbon ions in ion implantation apparatuses.